Key points
- Choose data according to workflows and obligations, not the number of files.
- Codes, units and data owners must be decided before the import.
- A test import and reconciliation of totals reduce surprises at go-live.
Inventory the data before cleaning it
List the files, exports and applications that contain customers, products, prices, suppliers, stock, orders and documents. Note who updates each source, how often and which decision depends on it.
Two files with the same name may have different owners and rules. Do not delete one because it looks old. First decide which source is accepted for each field and preserve the original separately.
Choose what is worth migrating
Active data and history needed for reporting or obligations should be considered separately. Current products, active customers, balances and open documents may matter more than every historical row.
Write inclusion and exclusion rules. A smaller import that can be checked is safer than a large archive that nobody can explain.
Clean and map the fields
Look for duplicates, different codes for the same product, mixed units, incomplete addresses and customers without a clear identity. Separate information that was placed in one cell and define the final format.
Create a map from old fields to new fields. For every transformation, write the rule, an example and the person who validates it. Important decisions should not remain hidden in a script the team cannot review.
Test the import as a process
Run an import on a sample that includes difficult cases. Check record counts, totals, links between orders and customers, stock and related documents.
Repeat the import to see whether the result is consistent and does not create duplicates. Test the operational reverse path too: can you identify the old row starting from the new code?
Plan the cutover day
Set the time when changes stop in the old systems, who makes the final export, who validates the import and what happens when a check fails. Keep the original copy and a rollback plan.
After go-live, review critical workflows and differences between old and new reports every day. Archive the sources with controlled access instead of letting them become a second operational database.
Relevant Webmate resources
Continue with guides, services and examples directly connected to the topic of this article.
Frequently asked questions
Do we need to migrate all historical data?
No. Migrate the data needed for operations, reporting and obligations. The rest can remain archived when it is accessible and verifiable.
Who validates migrated data?
Process owners and responsible operations or finance staff should validate the result, not only the technical supplier.
How do we prevent duplicates during import?
Set identifiers, matching rules and repeatable tests before the final import. Check totals after every run.
